Implementing Auto-Scaling with KEDA for a UK Healthcare Client


back to case studies

Implementing Auto-Scaling with KEDA for a UK Healthcare Client

Client Background

Our client, a prominent healthcare provider in the UK, faced challenges in managing the fluctuating demands on their digital services. To address this, we implemented an auto-scaling solution using Kubernetes-based Event Driven Autoscaling (KEDA). This case study details the journey from the discovery phase through to the successful implementation and outcomes.

Discovery Phase

Client Challenges

During the discovery phase, we conducted several workshops and interviews with the client’s IT and operations teams to understand their pain points and requirements. The key challenges identified were:

  1. Unpredictable Workloads: The client’s applications experienced significant fluctuations in demand, especially during peak hours and emergencies.
  2. Resource Inefficiency: Static scaling led to either over-provisioning, resulting in wasted resources, or under-provisioning, causing performance bottlenecks.
  3. Manual Scaling Limitations: The existing manual scaling processes were slow and error-prone, impacting the system's reliability and the team's productivity.

Requirements Gathering

Based on the challenges, we identified the following requirements for the auto-scaling solution:

  • Dynamic Scaling: Ability to automatically scale resources up or down based on real-time demand.
  • Cost Efficiency: Optimize resource usage to reduce operational costs.
  • Seamless Integration: The solution must integrate with existing infrastructure without major overhauls.
  • Reliability and Performance: Ensure consistent application performance and high availability.

Design Phase

Solution Architecture

We proposed an auto-scaling solution leveraging KEDA, an open-source project that provides event-driven auto-scaling for Kubernetes workloads. The architecture included:

  1. Kubernetes Cluster: The foundation for deploying and managing containerized applications.
  2. KEDA: To enable event-driven auto-scaling based on various event sources such as messaging queues, databases, and custom metrics.
  3. Monitoring and Metrics: Integration with Prometheus and Grafana for real-time monitoring and alerting.
  4. CI/CD Pipeline: Enhanced continuous integration and deployment pipelines to streamline updates and scaling policies.

Proof of Concept (PoC)

We developed a PoC to demonstrate the feasibility and benefits of KEDA. This involved:

  • Deploying a Sample Application: Simulating the client’s workload to test auto-scaling capabilities.
  • Configuring Scalers: Setting up KEDA scalers to respond to metrics such as CPU usage, queue length, and custom application metrics.
  • Testing: Running stress tests to observe auto-scaling behavior and performance improvements.

Implementation Phase

Setting Up the Environment

We began by setting up the Kubernetes environment and integrating KEDA. Key steps included:

  1. Cluster Configuration: Provisioning a Kubernetes cluster tailored to the client's needs, ensuring high availability and security.
  2. KEDA Deployment: Installing KEDA and configuring it with necessary permissions and access to metrics sources.
  3. Application Migration: Containerizing the client's applications and deploying them to the Kubernetes cluster.

Auto-Scaling Configuration

With the environment ready, we focused on configuring KEDA for the client’s specific use cases:

  1. Defining Metrics: Collaborating with the client to identify critical metrics that would drive scaling decisions.
  2. Scaler Configuration: Setting up KEDA scalers for various event sources like Azure Queue Storage, Prometheus metrics, and custom HTTP triggers.
  3. Policy Tuning: Fine-tuning scaling policies to balance performance and cost efficiency, including cooldown periods, scaling thresholds, and replica limits.

Testing and Optimization

Extensive Testing

We conducted extensive testing to ensure the solution met all performance and reliability requirements:

  • Load Testing: Simulating peak loads to observe auto-scaling responsiveness and application stability.
  • Failure Scenarios: Testing the system’s behavior during node failures, network issues, and other disruptions.
  • Performance Monitoring: Using Prometheus and Grafana to monitor real-time performance and resource usage, making adjustments as needed.

Feedback and Iteration

Throughout the testing phase, we maintained close communication with the client, gathering feedback and iterating on the configuration to address any issues and optimize performance further.

Deployment and Training


After successful testing and client approval, we rolled out the auto-scaling solution to production:

  • Staged Deployment: Gradual rollout to minimize risks, starting with non-critical workloads and expanding to critical services.
  • Monitoring and Support: Continuous monitoring during the initial weeks of deployment to quickly address any issues.

Client Training

To ensure the client’s team could effectively manage and leverage the new system, we provided comprehensive training:

  • Workshops and Documentation: Detailed training sessions and thorough documentation on managing KEDA and the Kubernetes environment.
  • Best Practices: Guidance on best practices for maintaining and scaling applications, monitoring performance, and troubleshooting.

Results and Benefits

Improved Performance and Availability

The auto-scaling solution significantly improved the performance and availability of the client’s applications, especially during peak times, ensuring reliable service delivery.

Cost Savings

By dynamically adjusting resources based on demand, the client achieved substantial cost savings, reducing over-provisioning and optimizing resource usage.

Enhanced Operational Efficiency

Automation of the scaling process reduced the manual workload on the IT team, allowing them to focus on more strategic initiatives.


The implementation of KEDA for auto-scaling has transformed the client’s ability to manage fluctuating demands efficiently. This case study underscores the importance of a tailored, well-executed solution in addressing specific industry challenges and achieving operational excellence in the healthcare sector.

Trusted by the world’s most innovative teams

coventry warwickshirerunningsburberryjupiterinvestmentsequity release supermarketrise interactiveibpLPCfantasywingsferoelectric coin cobluesky

Ready to Transform your Business?​

Harness the power of scalable and secure solutions designed to optimize your workflow, enhance collaboration, and drive unprecedented efficiency.

Whether you're looking to streamline processes, improve data management, or enhance overall performance, our tailored cloud services are the key to unlocking your company's full potential.